How does Naïve Bayes compare with kernel-based methods in high-dimensional spaces?

Naïve Bayes is faster but less expressive than kernel methods like SVM with RBF kernels.

Kernel methods transform data into higher-dimensional spaces to capture nonlinear relationships, while Naïve Bayes relies on probabilistic independence assumptions. NB is computationally cheaper (linear time), whereas kernel methods can model complex decision boundaries but at higher computational cost.